I have over 100 million rows in a PG table with geom type Point,
srid=4326. The points represent timestamped centroids of raster cells.
The data comes from sparse hourly 1km US grids; Only the cells with data,
which is less than 1%, but still a lot of points. The PG table allows for
a wealth of
Roger,
I don't know if this would help, but I recently came across a similar
problem which was fixed by adding the processing directive to the
layer:
PROCESSING RESAMPLE=NEAREST
This fixed a problem when zooming in very close on a raster layer.
Apparently without this, mapserver changes the